JP2019500191A - How to automatically determine an individual dichotomy index I <O - Google Patents

Abstract

  The object of the present invention is applicable to all individuals including very inactive people, such as elderly people or hospitalized people, without using the thresholding method applied by conventional algorithms To propose a method for automatically determining an individual's dichotomy index I <O by proposing means for automatically and specifically detecting activity and rest states based on the dichotomy index.

Description

  The present invention determines an individual's dichotomy index, an index that identifies alternating regularity of daytime activity and night rest, and their amplitude over 24 hours giving the possibility of providing a measure of circadian rhythm. Relates to the field of methods.

  Circadian disturbances that can be caused by working in shifts significantly increase the risk of cancer, especially breast, colon and prostate cancer. In addition, disruption of circadian rhythm of activity and rest measured by activity measurements is independent of known prognostic factors in patients with metastatic colorectal cancer, breast cancer, kidney cancer, ovarian cancer or lung cancer. Represents a negative prognostic factor from a survival perspective.

  Activity measurement is a non-invasive technique for measuring activity and rest cycle, and consists of sleep, wake rhythm, or circadian rhythm. To do this, an activator is used that records successive zero crossings of acceleration, which looks like a casing with at least one accelerometer.

  This type of device connected to the server that allows analysis of the collected data offers the possibility of reliably predicting activity and sleep duration that correlates well with that detected by the polysomnograph However, the uncertainty about falling asleep is greater than the uncertainty about awakening.

  Activometers are considered the “gold standard” for measuring activity and rest, but have severe pathology and / or very behavioral disabilities and therefore low physical activity ability Not ideal for certified individuals. This is in particular due to the fact that the activity level is considerably lower than that of healthy people, thereby making it very difficult to distinguish between the resting phase and the arousal phase, especially for sleep.

  Common methods for detecting activity in this type of individual are smoothing, logical combination or Tilmann, J. et al. “Algorithms for sleep-wake identification using activity: a comparative study and new results” in Journal of Sleep Research 2009, 18, or (1) pp85-98 Adapt the detection threshold by using a non-parametric algorithm.

  In particular, these are neural networks that give the best results. However, despite these algorithmic advances, the percentage of errors remains persistent in detecting a break between rest and awakening in hospitalized individuals.

"Algorithms for sleep-wake identification using activity: a comparable study and new results" Journal of Sleep Research 2009, 18, (1) pp85-98.

  Thus, the present invention can be applied to all individuals, including those with very low activity, such as elderly people or hospitalized people, without using the thresholding method applied by conventional algorithms. By suggesting a means for automatically and specifically detecting activity and rest based on a unique dichotomy index, the possibility of at least overcoming some of the drawbacks of the prior art is given, and the individual dichotomy index I <An object is to propose a method for automatically determining O.

ＮＢ_Ｃは、ステップＥ中に識別された横臥状態に位置する、個人の活動対時間を表す第１のＺＣＭ信号の値の数であり、その値は、ステップＥ中に識別された横臥状態の外側に位置する、個人の活動対時間を表す第１のＺＣＭ信号の値の中央値よりも大きい値であり、
ＮＢ_Ｌは、ステップＥ中に識別された横臥状態の外側に位置する、個人の活動（ＺＣＭ）に対する第１の信号の値の数である、ステップと
を含むことを特徴とする。 To this end, the present invention relates to a method for automatically determining a personal dichotomy index I <O from at least data from the system, the system comprising:
A temperature sensor that generates a data signal representing the temperature of the individual;
An accelerometer that simultaneously generates two data signals, a first “ZCM” signal representing personal activity over time, and a second “position X” signal representing personal inclination relative to a rising vertical line over time; A portable module comprising at least
A computer server comprising at least one processor, one memory and display means, configured to receive a data signal from a portable sensor and to perform the method,
The method has at least the following steps:
A. Receiving various data of the temperature, ZCM signal, and position X signal of the portable module by the processor and recording in memory;
B. If the module is not worn by an individual, the processor removes unavailable data from the module from memory;
C. Identifying anomalous data from the ZCM and position X signals of the accelerometer and records in their memory by the processor and resetting them to zero;
D. Converting the value of the accelerometer position X signal into a binary value (0 or 1) corresponding to the individual's arousal state (0) or recumbent state (1) by the processor and recording them in memory When,
E. Converting the value of the ZCM signal of the accelerometer into a binary value (0 or 1) corresponding to the individual's arousal or lying state by a processor and recording them in memory;
F. Over a period exceeding 180 minutes or over a period equal to 180 minutes, the processor compares the binary values of the accelerometer ZCM signal and the position X signal and the same binary for the accelerometer signal. Identifying at least one recumbent state of an individual constituted by a continuous sequence of values;
G. Calculating a bisection index by a processor according to the following equation and displaying it on the display means:

NB _C is positioned lying state identified in step E, the number of values of the first ZCM signal representing the individual activities versus time, the value of recumbent state identified in Step E A value that is greater than the median value of the first ZCM signal that represents the activity of the individual versus time, located outside;
NB _L is located outside the recumbent state identified in step E, the number of values of the first signal with respect to individual activities (ZCM), characterized in that it comprises a step.

  The bisection index I <O represents the percentage of activity per minute during the rest period (usually at night) and is less than the median value of activity away from the bed (usually daytime). This is a reliable indicator of circadian rhythms of activity and rest. This is calculated from the measurement of the number of accelerations per minute and the patient's position by a rib cage activator worn for at least 3 consecutive days. This is a non-invasive measurement method and is generally well accepted by patients.

  “ZCM” or zero crossing mode corresponds to the number of times the signal passes through zero for each time period.

  “Position X” corresponds to the value of the angle formed by the accelerometer relative to the rising vertical line when the accelerometer is worn on an individual's chest.

  According to a feature, the data signal comprises values obtained at regular intervals over a period of 24 hours.

  Another feature is that if the number of values of the data signal for the temperature sensor and the accelerometer are different, the processor uses the cubic spline polynomial interpolation to determine the data recorded in memory before step B. A normalization step is performed so that each value from the temperature signal is assigned to the value of the position X signal and the ZCM signal, and the value is recorded in the memory of the system.

  According to another feature, during step B, the processor determines that all values of the temperature data strictly below 30 ° C. and of the position X and ZCM signals temporally related to temperature values strictly below 30 ° C. Delete the value from memory.

According to another feature, during step C, the processor performs the following sub-steps on the entire position X and ZCM data:
a) selecting by a processor a first block of consecutive values in memory;
b) classifying block values in increasing order;
c) removing repeated block values to obtain a unique data set;
d) calculating the median of the unique data set;
e) resetting the values of the position X signal and the ZCM data signal to zero in the memory if the values of the position X signal and the ZCM data signal are greater than the median value;
f) In the next block of consecutive values, the steps repeating step a) to step e) are applied independently.

  According to another feature, a block of values is constituted by 31 values corresponding to a 31 minute period of the signal, and a median value is calculated over the first 10 data of the unique data set; A reset to zero is achieved by a subset of 5 data corresponding to 5 consecutive minutes, and if there are less than 2 values greater than the median, then the processor zeros the corresponding subset in memory. Reset to.

According to another feature, during step D, for the entire position X signal, the processor
The following substeps:
Categorizing values in increasing order;
Removing a repetition of values to obtain a unique data set;
Calculating the minimum median over the first 31 values of the unique data set and determining the minimum median and recording in memory;
The following substeps:
Classifying values in decreasing order; and
Removing a repetition of values to obtain a unique data set;
Calculating a maximum median value over the first 31 values of the unique data set;
Determining a threshold median corresponding to an average of the minimum median and maximum median and recording in memory;
In order to obtain the position X_binary data signal, the position X data signal is converted into memory by replacing a value strictly smaller than the threshold median with 0 and replacing a value greater than or equal to the threshold median with 1. Recording step;
When the value recorded in the memory of the position X_binary data signal by the processor is smoothed, and the set of data delimited by non-identical values corresponds to a period of less than 1 hour 30 minutes, the value of the data set is 0. The maximum median is determined and recorded in memory by performing the steps from 1 to 0 or 1 to 0.

According to another feature, during step E, for the entire ZCM signal, the processor
To obtain a new ZCM_time signal that varies between 0 and 1, normalize the ZCM value by dividing each ZCM value by the maximum value of the signal,
Apply a 10-point dispersion filter to the ZCM_Time signal, record the signal in memory,
Calculate the cumulative second average from the ZCM_time signal,
The following substeps using the processor:
Multiplied by factor 10 ⁴ to each value of the average secondary, calculating what rounded to the nearest integer,
Reading a window corresponding to 200 units at a pitch of 2 units;
Determining the number of points in each window, if the number of points found corresponds to a period longer than 180 minutes, the processor increments the “plateau” counter and sets the time index of the first point of the plateau Holding the time index of the last point in memory;
To obtain a ZCM_binary data signal, perform the steps of converting the ZCM data signal and recording it in memory by replacing the value located between the plateau time indices with 1 and replacing the other values with 0 By doing so, the plateau is searched for the cumulative second average data.

  According to another feature, during step F, the processor determines the ZCM_binary data signal and the position X_binary to determine a plateau time index corresponding to a lying state with an initial data value and a final data value of one. Perform a comparison with the data signal.

  According to another feature, during step G, the processor calculates data corresponding to the period of 1 hour before the start of the plateau and 1 hour after the start of the plateau, and before the end of the plateau, It does not take into account data corresponding to 1 hour and 1 hour after the end.

  Other features, details and advantages of the present invention will become apparent upon reading the following description with reference to the accompanying drawings.

It is a figure which shows an example of the value of a position X signal (step A). It is a figure which shows the same example of the value of the position X signal after deleting the data which cannot be used since the sensor is not worn (step B). It is a figure which shows the same example of the value of the position X signal after applying a filter with respect to abnormal data (step C). It is a figure which shows the same example of the position X_2 binary value corresponding to the same example of the value of the position X signal before smoothing (step D). It is a figure which shows the same example of the position X_binary value corresponding to the same example of the value of the position X signal after smoothing (step D). It is a figure which shows the example of ZCM_time value of the ZCM signal before applying a dispersion | distribution filter (step E). It is a figure which shows the same example of ZCM_time value of the ZCM signal after applying a dispersion | distribution filter (step E). It is a figure which shows the same example of ZCM_time value of the ZCM signal after applying a dispersion filter using a cumulative secondary average (step E). It is a figure which shows the ZCM_2 binary value after conversion of a ZCM signal (step E). It is a figure which shows a value required for calculation of a bisection index. FIG. 5 shows a readout window that allows determination of the presence of a plateau during step E.

A non-limiting example of an embodiment of a method for automatically determining a personal dichotomy index I <O according to the present invention will now be described. A temperature sensor that generates a data signal representing the individual's body temperature;
An accelerometer that simultaneously generates two data signals, a first ZCM signal representing personal activity over time, and a second position X signal representing a personal inclination relative to a rising vertical line over time; Use data from at least configured mobile modules.

  As an example, the accelerometer used can be an activity readout (or “ZCM” zero-crossing mode ZCM (zero-crossing mode) corresponding to the number of times the signal crosses zero for each period) or 1 minute It can be an “ADXL345” accelerometer that can generate a reading of every tilt (or “position X”). The temperature sensor is preferably an infrared type sensor configured to measure body temperature every 5 minutes.

  The portable module may be connected to the collector module via a wireless communication circuit via, for example, Bluetooth, WIFI, or GPRS connection. The collector module communicates signals to the server. This communication includes, for example, a processing unit such as a program that uses data that is temporarily stored in the memory of the processor, memory, and module and sent to the memory of the server via the communication circuit, such as a server. It can be GPRS, Bluetooth (registered trademark), WIFI, LIFI, infrared, wireless or wired, depending on the communication circuit of the computing resource.

  In some embodiments, the mobile module need not be connected to a server. In fact, a portable module can include processing resources such as processors, memory, and programs that use data from sensors and accelerometers temporarily stored in the module's memory, and can include personal resources. The method of automatically determining the bisection index I <O is executed. In other words, the functions of the mobile module and the server can be reintegrated in one mobile element.

  Thus, every 24 hours, the server, for example, uses a first signal corresponding to a continuation of a representative value of the individual's body temperature taken every 5 minutes over a 24 hour period (and of course more frequently used) Can be, but not necessarily appropriate if the change in body temperature is slow), for example, a second “corresponding to a continuous value representing an individual's activity acquired every minute over a period of 24 hours” ZCM ”signal and a third“ position X ”signal (FIG. 1) corresponding to a series of values representing, for example, an individual's inclination with respect to the rising vertical line acquired every minute over a period of 24 hours. Measurement values corresponding to the data signals are received.

  “24 hours” means a period of 24 hours, which can include a correction factor (which can be added or subtracted to this period) from a few minutes up to a maximum of about 1 hour. This correction factor may be calculated based on an individual's previous measurements, for example, to determine the duration of these cycles over one or more past cycles. This factor can also be determined based on a statistical estimate of the duration of the individual's future cycle, eg, based on past measurements and / or various prediction factors.

  To obtain the same number of temperature values as the “Position X” or “ZCM” value, the server performs a normalization step on the temperature data recorded in the memory using cubic spline polynomial interpolation. As a result, each temperature readout provides five values to obtain a temperature value for each “Position X” and “ZCM” value versus time.

  Subsequently, when the mobile module is not worn, a step of deleting data generated by the mobile module is executed by the server. For this purpose, the server includes a program that, during its execution, allows the identification of temperature values below 30 ° C. as well as its time index, ie the time reference for taking measurements. The algorithm deletes from memory the “ZCM” and “Position X” values that have temperature values below 30 ° C. and the same time index (FIG. 2).

  If the mobile module is worn upside down, a step of correcting the value from the “position X” tilt signal may be performed so that it is not detected with a negative angle value. For this reason, if the “Position X” value is greater than 90 °, the program executed by the server then subtracts the “Position X” value from 180 to convert: 180- “Position X” value Execute.

Subsequently, the server performs the steps of identifying abnormal data from the “ZCM” and “Position X” signals by the processor and resetting them to zero. To do this, the processor performs the following sub-steps for the entire “location X” and “ZCM” data:
a) selecting by a processor a first X block (e.g. 31) that is a continuous value corresponding to the duration X (e.g. 31 minutes) of an important measurement in memory;
b) classifying block values in increasing order;
c) removing repeated block values to obtain a unique data set;
d) calculating a median of the first values of the unique data set;
e) A block of X values (31) per Y sub-block (for example, 5) data corresponding to the Y measurement (each 5) is read, and a value less than Z <V greater than the median (for example, 2) ) Is present, resetting the Y sub-block (eg, 5) data to zero in memory;
f) In the next block of successive values, the steps repeating step a) to step e) are executed independently. (Figure 3).

  Following the step of removing outliers “ZCM” and “position X”, the server converts the values of the “position X” and “ZCM” signals to binary values 0 or 1. A binary value of 1 corresponds to a resting state and a binary value of 0 corresponds to an active state.

To do this, for the entire position X signal, the processor
The following substeps:
Categorizing values in increasing order;
Removing a repetition of values to obtain a unique data set;
Calculating the minimum median over the first X (31) values of the unique data set and determining the minimum median and recording in memory;
The following substeps:
Classifying values in decreasing order; and
Removing a repetition of values to obtain a unique data set;
Determining the maximum median by recording the maximum median over the first X (31) values of the unique data set, and recording in memory;
Determine the median threshold, which corresponds to the average of the minimum median and maximum median, and record it in memory;
To obtain the position X_binary data signal (FIG. 4), replace the value strictly below the median threshold with 0 and replace the value greater than or equal to the median with 1 with the “position X” data signal. Convert, record in memory,
If the value recorded in the memory of the position X_binary data signal is smoothed and the set of data delimited by non-identical values corresponds to a period of less than 1 hour 30 minutes, the value of the data set is from 0 to 1 And from 1 to 0 (FIG. 5).

  In other words, if the number of binary values included in the set is less than 90, then the binary values of the set are inverted.

To convert the “ZCM” value to a binary value, the processor performs the following operations over the entire “ZCM” signal:
Normalizing the “ZCM” values by dividing each “ZCM” value by the maximum value of the signal to obtain a new data set called ZCM_Time (FIG. 6) that varies between 0 and 1; ,
Applying the dispersion filter to the “ZCM_Time” data set over 10 points.

  Therefore, within “ZCM_Time” for each point where the time index i is modified, the dispersion filter will range from i to i + 9 (FIG. 7).

For example:
ZCM_Time [0] =
Calculation_Variance (ZCM_Time [0], ZCM_Time [1], ..., ZCM_Time [9])
ZCM_Time [1] =
Calculation_Variance (ZCM_Time [1], ZCM_Time [2], ..., ZCM_Time [10])
ZCM_Time [2] =
Calculation_Variance (ZCM_Time [2], ZCM_Time [3], ..., ZCM_Time [11])
The calculation of the ZCM_time signal, i.e. the cumulative second average RMS from each RMS point (of time index i), is the result of the previous RMS point (time index i-1) and the RMS result over two ZCM_time points. It is the sum (FIG. 8).

For example:
RMS [0] = Calculation_RMS (ZCM_Time [0], ZCM_Time [1])
RMS [1] = Calculation_RMS (ZCM_Time [1], ZCM_Time [2]) + RMS [0]
RMS [2] = Calculation_RMS (ZCM_Time [2], ZCM_Time [3]) + RMS [1]

In addition, the processor performs the following operations on the entire “ZCM” signal:
The step of determining the plateau for the cumulative second average data by performing the following sub-steps:
Multiplied by factor 10 ⁴ to each value of the average secondary, calculating what rounded to the nearest integer,
The step of reading the corresponding window at a pitch of 2 units, for example 200 units, in other words the window is moved from 2 to 2 giving the possibility of reading up to 200 values of the secondary average at the same time. That is, the window width is 200 units and its length is the size of the signal (FIG. 11), steps,
If the window contains at least 180 quadratic averages, the step of determining the presence of a plateau, the program executed by the processor then keeping a time index of the first and last point of the plateau in memory Performing a step and performing a step,
To obtain the “ZCM — Binary” data signal (FIG. 9), the ZCM data signal is converted into the memory by replacing the value located between the plateau time indices with 1 and replacing the other values with 0. Perform the steps to record and.

  Following these binary conversion steps, the server compares the binary values of the “ZCM” and “Position X” signals and is represented by a set of at least 180 binary values equal to 1. At least one rest state is identified.

  Finally, the server proceeds to display on a display means such as a screen by calculating a bisection index and applying the following equation:

ＮＢ_Ｃは、ステップＥ中に識別された横臥状態に位置する、個人の活動対時間を表す第１の「ＺＣＭ」信号の値の数であり、その値は、ステップＥ中に識別された横臥状態の外側に位置する、個人の活動対時間を表す第１の「ＺＣＭ」信号の値の中央値よりも大きい値であり、
ＮＢ_Ｌは、ステップＥ中に識別された横臥状態の外側に位置する、個人の活動（ＺＣＭ）を表す第１のＺＣＭ信号の値の数である。

NB _C is positioned lying state identified in step E, the number of values of the first "ZCM" signal representing the individual activities versus time, the value, lying identified in Step E A value greater than the median value of the first “ZCM” signal representing the individual's activity versus time, located outside the state;
NB _L is the number of values of the first ZCM signal representing the personal activity (ZCM) located outside the recumbent state identified during step E.

In order to do this, the server executes the program on its processor by the following sub-steps:
Identifying a set of “ZCM” values that are stored in memory under the name “ZCM_C” and whose points are included in the plateau;
Identifying a set of “ZCM” values that are stored in memory under the name “ZCM_L” and whose points are on either side of the plateau;
Calculating a median “Med_L” corresponding to the ZCM_L value;
Counting the number of points NB_C that exceeds the median value “Med_L” in “ZCM_C” and the number of points NB_L in “ZCM_L”;
And solving the bisection index equation.

  In a preferred embodiment, in order to improve accuracy, the processor calculates data for the bisection index for 1 hour before the start of the plateau and 1 hour after the start of the plateau, and 1 hour before the end of the plateau. And data corresponding to one hour after the end is not taken into account. In other words, ZCM_C corresponds to the set of ZCM values whose points are included in the plateau, except for the first and last hour of the plateau, and ZCM_L excludes one hour before the plateau and one hour immediately after the plateau. The point corresponds to a set of ZCM values on both sides of the plateau.

  Thus, in this particular embodiment, to calculate the dichotomy index, the latter is only calculated over a 20 hour period per 24 hour slice (FIG. 10).

  According to another embodiment, if this is required to calculate the median, a “ZCM” or “position X” signal is used to obtain a sufficient number of values to perform the median calculation. A set of 16 values identical to the last value of can be added to the end of the signal.

  Furthermore, if the dichotomy index is greater than 97%, this means that the individual is well asleep and that person's circadian rhythm is not disturbed.

  A dichotomy index of less than 97% represents a poor prognostic factor in terms of survival, especially in hospitalized individuals, especially for cancer patients.

  Obtaining this result gives the possibility to know the circadian rhythm of an individual suffering from cancer and enables optimization of time-controlled administration of anticancer drugs, thereby improving the resistance and efficiency of anticancer drugs. It is possible to improve.

  Such a method according to the present invention offers the possibility of helping the doctor with the aim of refining the optimal time and treatment plan for each individual.

  In addition, all of the above variables can be dynamic variables defined based on measured data and / or statistical estimates (such as those described above for 24 hours). Thus, correction factors can be added or subtracted from them. Such a correction factor can take into account individual variations between patients for a given variation, eg, as described herein.

  The description of the invention and the related drawings are not provided to limit the scope of the invention, but merely illustrate selected embodiments. Those skilled in the art will understand that the technical features of a given embodiment are not otherwise stated unless it is actually stated that the opposite is true, or it is clear that these features are incompatible. It can actually be combined with the features of certain other embodiments. Furthermore, the technical features described in a particular given embodiment can be separated from the other features of this embodiment, unless the contrary is explicitly stated.

  It will be apparent to those skilled in the art that the present invention allows embodiments under many other specific forms without departing from the scope defined by the appended claims. They should be considered exemplary and the invention should not be limited to the given details given above.

Claims (13)

A method for automatically determining a personal dichotomy index I <O from at least data from a system, the system comprising:
i. A temperature sensor that generates a data signal representing the temperature of the individual;
ii. An accelerometer that simultaneously generates two data signals: a first “ZCM” signal representing personal activity over time, and a second “position X” signal representing personal inclination relative to a rising vertical line over time A mobile module including and
A computing resource including at least one processor and at least one memory, receiving a data signal from a portable sensor and configured to perform the method,
At least the following steps are performed by the processor by the method:
A. Receiving various data of the temperature, ZCM signal, and position X signal of the portable module and recording them in memory;
B. Deleting the unavailable data from the module from memory if the module is not worn by an individual;
C. Identifying anomalous data from the ZCM and position X signals of the accelerometer and records in their memory by the processor and resetting to zero;
D. Converting the value of the accelerometer position X signal into a binary value (0 or 1) corresponding to the individual's arousal state (0) or recumbent state (1) and recording them in memory;
E. Converting the value of the ZCM signal of the accelerometer into binary values corresponding to an individual's arousal or lying state and recording them in memory;
F. Compare binary values of accelerometer ZCM signal and position X signal over a period of more than 180 minutes, or over a period of 180 minutes, and the same binary value breaks for the accelerometer signal Identifying at least one recumbent state of an individual constituted by no succession;
G. Calculating a bisection index according to the following equation and displaying it on the display means:

NB _C is positioned lying state identified in step E, the number of values of the first ZCM signal representing the individual activities versus time, the value of recumbent state identified in Step E A value that is greater than the median value of the first ZCM signal that represents the activity of the individual versus time, located outside;
NB _L is located outside the recumbent state identified in step E, the number of values of the first signal with respect to individual activities (ZCM), a method which comprises the steps.   The method for automatically determining an individual's dichotomy index I <O according to claim 1, characterized in that computational resources are integrated in a portable module for performing said method.   2. The method for automatically determining a personal dichotomy index I <O according to claim 1, characterized in that the computational resources are integrated in a different computer server than the portable module.   4. A method for automatically determining a personal dichotomy index I <O according to claim 3, characterized in that the computer server comprises display means for displaying the result of the calculation.   3. Automatically determining the individual dichotomy index I <O according to claim 1 or 2, characterized in that the data signal comprises values obtained at regular intervals over a period of 24 hours. Method.   If the number of data signal values for the temperature sensor and accelerometer are different, the processor performs a normalization step on the data recorded in memory before step B by using cubic spline polynomial interpolation As a result, each value from the temperature signal is assigned to the value of the position X signal and the ZCM signal, and the value is recorded in the memory of the system. A method for automatically determining the individual dichotomy index I <O described in the paragraph.   During step B, the processor deletes all values of temperature data strictly below 30 ° C., and values of position X and ZCM signals temporally related to temperature values strictly below 30 ° C. from memory. 5. A method for automatically determining an individual's dichotomy index I <O according to any one of claims 1 to 4, characterized by: During step C, the processor performs the following substeps for the entire position X and ZCM data:
a) selecting by a processor a first block of consecutive values in memory;
b) classifying block values in increasing order;
c) removing repeated block values to obtain a unique data set;
d) calculating the median of the unique data set;
e) resetting the values of the position X signal and the ZCM data signal to zero in the memory if the values of the position X signal and the ZCM data signal are greater than the median value;
f) the step of repeating steps a) to e) in the next block of successive values, and independently applying the individual bisection index I <O of claim 5 How to determine automatically. A block of values consists of 31 values corresponding to a 31 minute period of the signal,
A median is calculated over the first 10 data of the unique data set;
A reset to zero is achieved by a subset of 5 data corresponding to 5 consecutive minutes, and if there are less than 2 values greater than the median, then the processor zeros the corresponding subset in memory. 7. The method of automatically determining an individual's dichotomy index I <O according to claim 6, characterized in that it is reset to. During step D, for the entire position X signal, the processor
The following substeps:
Categorizing values in increasing order;
Removing a repetition of values to obtain a unique data set;
Calculating the minimum median over the first 31 values of the unique data set and determining the minimum median and recording in memory;
The following substeps:
Classifying values in decreasing order; and
Removing a repetition of values to obtain a unique data set;
Calculating a maximum median value over the first 31 values of the unique data set;
Determining a threshold median corresponding to an average of the minimum median and maximum median and recording in memory;
In order to obtain the position X_binary data signal, the position X data signal is converted into memory by replacing a value strictly smaller than the threshold median with 0 and replacing a value greater than or equal to the threshold median with 1. Recording step;
When the value recorded in the memory of the position X_binary data signal by the processor is smoothed, and the set of data delimited by non-identical values corresponds to a period of less than 1 hour 30 minutes, the value of the data set is 0. The personal dichotomy index according to claim 7, characterized in that the maximum median is determined and recorded in memory by performing the steps reversed from 1 to 1 or from 1 to 0 A method of automatically determining I <O. During step E, for the entire ZCM signal, the processor
To obtain a new ZCM_time signal that varies between 0 and 1, normalize the ZCM value by dividing each ZCM value by the maximum value of the signal,
Apply a 10-point dispersion filter to the ZCM_Time signal, record the signal in memory,
Calculate the cumulative second average from the ZCM_time signal,
The following substeps using the processor:
Multiplied by factor 10 ⁴ to each value of the average secondary, calculating what rounded to the nearest integer,
Reading a window corresponding to 200 units at a pitch of 2 units;
Determining the number of points in each window, if the number of points found corresponds to a period longer than 180 minutes, the processor increments the “plateau” counter and sets the time index of the first point of the plateau Holding the time index of the last point in memory;
To obtain a ZCM_binary data signal, perform the steps of converting the ZCM data signal and recording it in memory by replacing the value located between the plateau time indices with 1 and replacing the other values with 0 9. The method of automatically determining an individual bisection index I <O according to claim 8, wherein a plateau is searched for cumulative second average data.   During step F, the processor performs a comparison between the ZCM_binary data signal and the position X_binary data signal to determine a plateau time index corresponding to a lying state with an initial data value and a final data value of 1. 10. The method for automatically determining an individual's dichotomy index I <O according to claim 9, characterized in that:   During step G, the processor calculates data for the bisection index, corresponding to a period of 1 hour before the start of the plateau and 1 hour after the start of the plateau, and 1 hour before the end of the plateau and 1 after the end of the plateau. 11. A method for automatically determining an individual's bisection index I <O according to claim 10, characterized in that it does not take into account data corresponding to a period of time.

Images